Natural purines with a nitrogen oxidized to an N-oxide are oncogens. Possibly they arise endogenously, and chemical oncogenesis may not be limited to exogenous non-biological compounds. Structure-activity relationships will continue to yield useful information. Studies of 3-hydroxyguanine and 3-hydroxyxanthine, the most potent oncogens, are emphasized. Each is metabolized to a "proximate" oncogen. Those each lead to two types of chemically reactive intermediates: an electrophilic carbonium ion and a radical with hydrogen abstracting activity. To investigate the possible role radicals may play in oncogenesis, the structural features favoring each type of chemically reactive intermediate are sought by correlations of the behaviors of chemical models of the proximate oncogen(s) and studies of photochemical formation of radicals from the solids, or from solutions of, a variety of purine N-oxide derivatives. Integration of information on oncogenicities, on in vivo and enzymic metabolism, and on chemical properties and reaction mechanisms, will attempt to clarify the structural features necessary for specific types of metabolism and reactivities. Definition of the "ultimate" oncogen, and clarification of its critical sites of action and effects on metabolic controls are an ultimate goal. Fates of the weakly oncogenic adenine 1-oxide, and of certain 3-hydroxyxanthine derivatives, must be clarified, since the fates which are known do not relate them to the fates observed for 3-hydroxyxanthine.